Detection of SST fronts from high-resolution model and its preliminary results in the South China Sea
AbstractA frontal detection algorithm is developed with the capability of detecting significant frontal segments of sea surface temperature (SST) in the high resolution South China Sea Operational Forecasting System (SCSOFS). In order to effectively obtain frontal information, a gradient-based Canny edge detection algorithm is improved with post-processing designed for high resolution numerical model, aiming at extracting primary ocean fronts, meanwhile ensuring the balance of frontal continuity and positioning accuracy. Metrics of frontal probability and strength are used to measure the robustness of the results in terms of mean state and seasonal variability of frontal activities in the South China Sea (SCS). Most fronts are found in the nearshore and forming a strip shape extending from the Taiwan Strait to the coast of Vietnam. The SCSOFS is found to reproduce strong seasonal signals dominating the variability of the frontal strength and occurrence probability in the SCS. We implement the algorithm on the daily-averaged SST derived from two other SST analyses for inter-comparison in the SCS.